import pandas as pd 
import numpy as np
import matplotlib.pyplot as plt 
import seaborn as sns 
import warnings
from sklearn.exceptions import FitFailedWarning
warnings.filterwarnings("ignore", category=FitFailedWarning)


from sklearn.tree import DecisionTreeClassifier
from sklearn.impute import KNNImputer
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score
from imblearn.over_sampling import SMOTE
from imblearn.under_sampling import RandomUnderSampler

from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.linear_model import LogisticRegression
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import confusion_matrix, classification_report
from xgboost import XGBClassifier
import pandas as pd
from sklearn.feature_selection import SelectKBest, f_classif

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from sklearn.model_selection import GridSearchCV
from tensorflow.keras.optimizers import Adam
from sklearn.ensemble import BaggingClassifier



from sklearn.ensemble import RandomForestClassifier
# Set display options to show all rows and columns
pd.set_option('display.max_rows', None)
pd.set_option('display.max_columns', None)
pd.set_option('display.width', None)
pd.set_option('display.max_colwidth', None)

# Your code here

# Reset display options to default
pd.reset_option('display.max_rows')
pd.reset_option('display.max_columns')
pd.reset_option('display.width')
pd.reset_option('display.max_colwidth')

df = pd.read_csv(r"C:\Users\Admin\Downloads\archive (24)\water_potability.csv")

df

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3276 entries, 0 to 3275
Data columns (total 10 columns):
 #   Column           Non-Null Count  Dtype  
---  ------           --------------  -----  
 0   ph               2785 non-null   float64
 1   Hardness         3276 non-null   float64
 2   Solids           3276 non-null   float64
 3   Chloramines      3276 non-null   float64
 4   Sulfate          2495 non-null   float64
 5   Conductivity     3276 non-null   float64
 6   Organic_carbon   3276 non-null   float64
 7   Trihalomethanes  3114 non-null   float64
 8   Turbidity        3276 non-null   float64
 9   Potability       3276 non-null   int64  
dtypes: float64(9), int64(1)
memory usage: 256.1 KB

print("Missing total")
print(df.isna().sum())
print('---------')
print('Missing percentages')
print(df.isna().sum() / df.shape[0])
print('---------')
missing = []
columns = df.columns
for i, v in enumerate(df.isna().sum()):
    if v > 0:
        missing.append(columns[i])
print(missing)

Missing total
ph                 491
Hardness             0
Solids               0
Chloramines          0
Sulfate            781
Conductivity         0
Organic_carbon       0
Trihalomethanes    162
Turbidity            0
Potability           0
dtype: int64
---------
Missing percentages
ph                 0.149878
Hardness           0.000000
Solids             0.000000
Chloramines        0.000000
Sulfate            0.238400
Conductivity       0.000000
Organic_carbon     0.000000
Trihalomethanes    0.049451
Turbidity          0.000000
Potability         0.000000
dtype: float64
---------
['ph', 'Sulfate', 'Trihalomethanes']

df.columns

Index(['ph', 'Hardness', 'Solids', 'Chloramines', 'Sulfate', 'Conductivity',
       'Organic_carbon', 'Trihalomethanes', 'Turbidity', 'Potability'],
      dtype='object')

X = df.drop('Potability', axis=1)
imputer = KNNImputer()
X_imputed = imputer.fit_transform(X)
df_ = pd.DataFrame(X_imputed, columns=X.columns)

y = df['Potability']
df = pd.concat([df_, y], axis=1)
df

for col in df.columns[:-1]:
    plt.figure(figsize=(8, 6))
    sns.histplot(data=df,x=col, hue='Potability', kde=True, alpha=0.7)
    plt.show()

print('Feature with high skewness:')
for i, v in enumerate(df.skew()):
    if v > 0.5 or v < -0.5:
        print(columns[i], ':', v)

Feature with high skewness:
Solids : 0.6216344855169127

sns.pairplot(df)

C:\Users\Admin\anaconda3\Lib\site-packages\seaborn\axisgrid.py:118: UserWarning: The figure layout has changed to tight
  self._figure.tight_layout(*args, **kwargs)

<seaborn.axisgrid.PairGrid at 0x1fb9e911510>

correlation = df.corr()
plt.figure(figsize=(12, 8))
sns.heatmap(correlation, annot=True, lw=1)

<Axes: >

df.columns

Index(['ph', 'Hardness', 'Solids', 'Chloramines', 'Sulfate', 'Conductivity',
       'Organic_carbon', 'Trihalomethanes', 'Turbidity', 'Potability'],
      dtype='object')

df['ph_level'] = pd.cut(df['ph'], bins=[0, 6.9, 7.1, 14], labels=['acidic', 'neutral', 'alkaline'], include_lowest=True)
df['tds_level'] = pd.cut(df['Solids'], bins=[0, 500, 1000, float('inf')], labels=['low', 'moderate', 'high'], include_lowest=True)
df['chloramines_level'] = pd.cut(df['Chloramines'],  bins=[0, 4, float('inf')], labels=['low', 'high'], include_lowest=True)
df['trihalomethanes_level'] = pd.cut(df['Trihalomethanes'], bins=[0, 80, float('inf')], labels=['safe', 'unsafe'], include_lowest=True)

ordinal_mapping = {
    'acidic': 0,
    'neutral': 1,
    'alkaline': 2,
    'low': 0,
    'moderate': 1,
    'high': 2,
    'safe': 0,
    'unsafe': 1,
    'low': 0,
    'high': 1
}

df['ph_level'] = df['ph_level'].map(ordinal_mapping)
df['tds_level'] = df['tds_level'].map(ordinal_mapping)
df['chloramines_level'] = df['chloramines_level'].map(ordinal_mapping)
df['trihalomethanes_level'] = df['trihalomethanes_level'].map(ordinal_mapping)

df

y

0       0
1       0
2       0
3       0
4       0
       ..
3271    1
3272    1
3273    1
3274    1
3275    1
Name: Potability, Length: 3276, dtype: int64

df

correlation = df.corr()
plt.figure(figsize=(12, 8))
sns.heatmap(correlation, annot=True, lw=1)

<Axes: >

df['Potability'].value_counts()

Potability
0    1998
1    1278
Name: count, dtype: int64

X, y = df.drop('Potability', axis=1), df['Potability']
oversampler = SMOTE(sampling_strategy='auto', random_state=42)
undersampler = RandomUnderSampler(sampling_strategy='auto', random_state=42)

X_resampled, y_resampled = oversampler.fit_resample(X, y)
X_resampled, y_resampled = undersampler.fit_resample(X_resampled, y_resampled)
print(pd.Series(y_resampled).value_counts())

Potability
0    1998
1    1998
Name: count, dtype: int64

df.Potability.value_counts()

Potability
0    1998
1    1278
Name: count, dtype: int64

df_copy = df.copy()
df_copy['Potability'] = y_resampled

df_copy

correlation_after_resample = df_copy.corr()
plt.figure(figsize=(12, 8))
sns.heatmap(correlation_after_resample, annot=True, lw=1)

<Axes: >

X, y = df.drop('Potability', axis=1), df['Potability']
X_copy, y_copy = df_copy.drop('Potability', axis=1), df_copy['Potability']
X_train, X_test, y_train, y_test =train_test_split(X,y, test_size=0.2, random_state=42)
X_train_copy, X_test_copy, y_train_copy, y_test_copy = train_test_split(X_copy, y_copy, test_size=0.2, random_state=42)

scale = StandardScaler()
X_train = scale.fit_transform(X_train)
X_test = scale.transform(X_test)
X_train_copy = scale.fit_transform(X_train_copy)
X_test_copy = scale.transform(X_test_copy)
                                                                       
lr = LogisticRegression()
param_grid = {
    'C': [0.01, 0.1, 1, 10, 100],
    'solver': ['liblinear', 'lbfgs', 'saga'],
    'max_iter': [100, 1000, 10000]
}

grid_base_line_lr = GridSearchCV(estimator=lr, param_grid=param_grid, cv=5, verbose=2, n_jobs=-1, error_score='raise')
grid_base_line_lr.fit(X_train, y_train)
y_pred_lr = grid_base_line_lr.predict(X_test)
print(classification_report(y_test, y_pred_lr))

grid_base_line_lr.fit(X_train_copy, y_train_copy)
y_pred_lr_copy = grid_base_line_lr.predict(X_test_copy)
print(classification_report(y_test_copy, y_pred_lr_copy))

Fitting 5 folds for each of 45 candidates, totalling 225 fits

C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

              precision    recall  f1-score   support

           0       0.63      1.00      0.77       412
           1       0.00      0.00      0.00       244

    accuracy                           0.63       656
   macro avg       0.31      0.50      0.39       656
weighted avg       0.39      0.63      0.48       656

Fitting 5 folds for each of 45 candidates, totalling 225 fits
              precision    recall  f1-score   support

           0       0.63      1.00      0.77       415
           1       0.00      0.00      0.00       241

    accuracy                           0.63       656
   macro avg       0.32      0.50      0.39       656
weighted avg       0.40      0.63      0.49       656

df

df_copy

df

fig, axs = plt.subplots(7, 2, figsize=(12, 24))
indices = 0
axs = axs.flatten()
for i in df_copy.columns:
    sns.boxplot(df[i], ax=axs[indices])
    axs[indices].set_xlabel(i)  # Set x-label
    indices += 1
plt.tight_layout()  # Adjust the layout
plt.show()

def remove_outliers(df, columns):
    for column in columns:
        q1 = df[column].quantile(0.25)
        q3 = df[column].quantile(0.75)
        iqr = q3 - q1
        lower = q1 - 1.5*iqr
        upper = q3 + 1.5*iqr
        df = df[(df[column] >= lower) & (df[column] <= upper )]
    return df

columns = ['ph', 'Solids', 'Chloramines', 'Sulfate', 'Conductivity', 'Organic_carbon', 'Trihalomethanes', 'Turbidity']
df_cleaned = remove_outliers(df, columns)

df_cleaned

param_grid = {
    'penalty': ['l1', 'l2'],
    'C': [0.01, 0.1, 1, 10, 100],
    'solver': ['liblinear', 'lbfgs', 'saga'],
    'max_iter': [100, 1000, 10000]
}
X, y = df_cleaned.drop('Potability', axis=1), df_cleaned['Potability']
X_train, X_test, y_train, y_test = train_test_split(X,y, test_size=0.2, random_state=42)
scale = StandardScaler()
X_train = scale.fit_transform(X_train)
X_test = scale.transform(X_test)
lr = LogisticRegression()
grid_remove_outliers = GridSearchCV(estimator=lr, param_grid=param_grid, cv=5, verbose=2, n_jobs=-1)
grid_remove_outliers.fit(X_train, y_train)
y_pred = grid_remove_outliers.predict(X_test)
print(accuracy_score(y_test,y_pred))
print(classification_report(y_test,y_pred))

Fitting 5 folds for each of 90 candidates, totalling 450 fits
0.6219931271477663
              precision    recall  f1-score   support

           0       0.62      1.00      0.77       362
           1       0.00      0.00      0.00       220

    accuracy                           0.62       582
   macro avg       0.31      0.50      0.38       582
weighted avg       0.39      0.62      0.48       582

C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\model_selection\_search.py:952: UserWarning: One or more of the test scores are non-finite: [0.6177987         nan 0.6177987  0.6177987  0.6177987  0.6177987
 0.6177987         nan 0.6177987  0.6177987  0.6177987  0.6177987
 0.6177987         nan 0.6177987  0.6177987  0.6177987  0.6177987
 0.6177987         nan 0.6177987  0.61822881 0.61822881 0.61822881
 0.6177987         nan 0.6177987  0.61822881 0.61822881 0.61822881
 0.6177987         nan 0.6177987  0.61822881 0.61822881 0.61822881
 0.6177987         nan 0.6177987  0.61822881 0.61822881 0.61822881
 0.6177987         nan 0.6177987  0.61822881 0.61822881 0.61822881
 0.6177987         nan 0.6177987  0.61822881 0.61822881 0.61822881
 0.61822881        nan 0.61822881 0.61822881 0.61822881 0.61822881
 0.61822881        nan 0.61822881 0.61822881 0.61822881 0.61822881
 0.61822881        nan 0.61822881 0.61822881 0.61822881 0.61822881
 0.61822881        nan 0.61822881 0.61822881 0.61822881 0.61822881
 0.61822881        nan 0.61822881 0.61822881 0.61822881 0.61822881
 0.61822881        nan 0.61822881 0.61822881 0.61822881 0.61822881]
  warnings.warn(
C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))
C:\Users\Admin\anaconda3\Lib\site-packages\sklearn\metrics\_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.
  _warn_prf(average, modifier, msg_start, len(result))

df

xg = XGBClassifier()
param_grid_xg = {
    'learning_rate': [0.01, 0.1],
    'n_estimators': [100, 500],
    'max_depth': [3, 5, 7],
    'gamma': [0.0, 0.1, 0.2],
    'subsample': [0.6, 0.8, 1.0],
    'colsample_bytree': [0.6, 0.8, 1.0],
}
grid_xg = GridSearchCV(estimator=xg, param_grid=param_grid_xg, cv=5, verbose=2, n_jobs=-1)
grid_xg.fit(X_train, y_train)
y_pred_xg = grid_xg.predict(X_test)
accuracy_xg = accuracy_score(y_test, y_pred_xg)
print(accuracy_xg)
print(classification_report(y_test, y_pred_xg))

Fitting 5 folds for each of 324 candidates, totalling 1620 fits
0.6632302405498282
              precision    recall  f1-score   support

           0       0.66      0.96      0.78       362
           1       0.71      0.18      0.29       220

    accuracy                           0.66       582
   macro avg       0.69      0.57      0.53       582
weighted avg       0.68      0.66      0.59       582

df.Potability.value_counts()

Potability
0    1998
1    1278
Name: count, dtype: int64

df_remove_balance = pd.concat([X_resampled, y_resampled], axis=1)

df_remove_balance

df_remove_balance.isna().sum()

ph                         0
Hardness                   0
Solids                     0
Chloramines                0
Sulfate                    0
Conductivity               0
Organic_carbon             0
Trihalomethanes            0
Turbidity                  0
ph_level                 398
tds_level                  0
chloramines_level         42
trihalomethanes_level    220
Potability                 0
dtype: int64

X = df_remove_balance.drop('Potability', axis=1)
imputer = KNNImputer()
X_imputed = imputer.fit_transform(X)
df_ = pd.DataFrame(X_imputed, columns=X.columns)

df_

df_remove_balance = pd.concat([df_, y_resampled], axis=1)

df_remove_balance

X, y = df_remove_balance.drop('Potability', axis=1), df_remove_balance['Potability']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
xg = XGBClassifier()
param_grid_xg = {
    'learning_rate': [0.01, 0.1],
    'n_estimators': [100, 500],
    'max_depth': [3, 5, 7],
    'gamma': [0.0, 0.1, 0.2],
    'subsample': [0.6, 0.8, 1.0],
    'colsample_bytree': [0.6, 0.8, 1.0],
}
grid_xg = GridSearchCV(estimator=xg, param_grid=param_grid_xg, cv=5, verbose=2, n_jobs=-1)
grid_xg.fit(X_train, y_train)
y_pred_xg = grid_xg.predict(X_test)
accuracy_xg = accuracy_score(y_test, y_pred_xg)
print(accuracy_xg)
print(classification_report(y_test, y_pred_xg))

Fitting 5 folds for each of 324 candidates, totalling 1620 fits
0.69
              precision    recall  f1-score   support

           0       0.69      0.73      0.71       415
           1       0.69      0.65      0.67       385

    accuracy                           0.69       800
   macro avg       0.69      0.69      0.69       800
weighted avg       0.69      0.69      0.69       800

X, y = df_remove_balance.drop('Potability', axis=1), df_remove_balance['Potability']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
scale = StandardScaler()
X_train = scale.fit_transform(X_train)
X_test = scale.transform(X_test)
xg = XGBClassifier()
param_grid_xg = {
    'learning_rate': [0.01, 0.1],
    'n_estimators': [100, 500],
    'max_depth': [3, 5, 7],
    'gamma': [0.0, 0.1, 0.2],
    'subsample': [0.6, 0.8, 1.0],
    'colsample_bytree': [0.6, 0.8, 1.0],
}
grid_xg = GridSearchCV(estimator=xg, param_grid=param_grid_xg, cv=5, verbose=2, n_jobs=-1)
grid_xg.fit(X_train, y_train)
y_pred_xg = grid_xg.predict(X_test)
accuracy_xg = accuracy_score(y_test, y_pred_xg)
print(accuracy_xg)
print(classification_report(y_test, y_pred_xg))

Fitting 5 folds for each of 324 candidates, totalling 1620 fits
0.69625
              precision    recall  f1-score   support

           0       0.70      0.73      0.71       415
           1       0.69      0.66      0.68       385

    accuracy                           0.70       800
   macro avg       0.70      0.70      0.70       800
weighted avg       0.70      0.70      0.70       800

df_remove_balance

df_remove_balance.Potability.value_counts()

Potability
0    1998
1    1998
Name: count, dtype: int64

X, y = df_remove_balance.drop('Potability', axis=1), df_remove_balance['Potability']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

k = 10

selector = SelectKBest(score_func=f_classif, k=k)
X_train_selected = selector.fit_transform(X_train, y_train)
X_test_selected = selector.transform(X_test)

selected_features_indices = selector.get_support(indices=True)

selected_features = X.columns[selected_features_indices]
print('Selected features: ', selected_features)

Selected features:  Index(['Hardness', 'Solids', 'Chloramines', 'Sulfate', 'Conductivity',
       'Organic_carbon', 'ph_level', 'tds_level', 'chloramines_level',
       'trihalomethanes_level'],
      dtype='object')

xg = XGBClassifier()
param_grid_xg = {
    'learning_rate': [0.01, 0.1],
    'n_estimators': [100, 500],
    'max_depth': [3, 5, 7],
    'gamma': [0.0, 0.1, 0.2],
    'subsample': [0.6, 0.8, 1.0],
    'colsample_bytree': [0.6, 0.8, 1.0],
}
grid_xg = GridSearchCV(estimator=xg, param_grid=param_grid_xg, cv=5, verbose=2, n_jobs=-1)
grid_xg.fit(X_train_selected, y_train)
y_pred_kbest = grid_xg.predict(X_test_selected)
accuracy_xg_kbest = accuracy_score(y_test, y_pred_kbest)
print(accuracy_xg_kbest)
print(classification_report(y_test, y_pred_kbest))

Fitting 5 folds for each of 324 candidates, totalling 1620 fits
0.6875
              precision    recall  f1-score   support

           0       0.68      0.76      0.72       415
           1       0.70      0.61      0.65       385

    accuracy                           0.69       800
   macro avg       0.69      0.68      0.68       800
weighted avg       0.69      0.69      0.69       800

rfc = RandomForestClassifier()
param_grid_rfc = {
    'n_estimators': [100, 200, 300],
    'max_depth': [None, 10, 20],
    'min_samples_split': [2, 5, 10],
    'min_samples_leaf': [1, 2, 4],
}

grid_rfc = GridSearchCV(estimator=rfc, param_grid=param_grid_rfc, cv=5, verbose=2, n_jobs=-1)
grid_rfc.fit(X_train_selected, y_train)
y_pred_kbest = grid_rfc.predict(X_test_selected)
accuracy_rfc_kbest = accuracy_score(y_test, y_pred_kbest)
print(accuracy_rfc_kbest)
print(classification_report(y_test, y_pred_kbest))

Fitting 5 folds for each of 81 candidates, totalling 405 fits
0.70625
              precision    recall  f1-score   support

           0       0.71      0.74      0.72       415
           1       0.70      0.67      0.69       385

    accuracy                           0.71       800
   macro avg       0.71      0.70      0.71       800
weighted avg       0.71      0.71      0.71       800

X, y = df_remove_balance.drop('Potability', axis=1), df_remove_balance['Potability']
X_train, X_test, y_train, y_test =train_test_split(X,y, test_size=0.2, random_state=42)
scale = StandardScaler()
X_train = scale.fit_transform(X_train)
X_test = scale.transform(X_test)
rfc = RandomForestClassifier()
param_grid_rfc = {
    'n_estimators': [100, 200, 300],
    'max_depth': [None, 10, 20],
    'min_samples_split': [2, 5, 10],
    'min_samples_leaf': [1, 2, 4],
}

grid_rfc = GridSearchCV(estimator=rfc, param_grid=param_grid_rfc, cv=5, verbose=2, n_jobs=-1)
grid_rfc.fit(X_train, y_train)
y_pred = grid_rfc.predict(X_test)
accuracy_rfc = accuracy_score(y_test, y_pred)
print(accuracy_rfc)
print(classification_report(y_test, y_pred))

Fitting 5 folds for each of 81 candidates, totalling 405 fits
0.73375
              precision    recall  f1-score   support

           0       0.73      0.76      0.75       415
           1       0.73      0.70      0.72       385

    accuracy                           0.73       800
   macro avg       0.73      0.73      0.73       800
weighted avg       0.73      0.73      0.73       800

df_remove_balance

df_remove_balance.skew()

ph                        0.034198
Hardness                 -0.016929
Solids                    0.615021
Chloramines              -0.027961
Sulfate                  -0.058976
Conductivity              0.276901
Organic_carbon            0.017890
Trihalomethanes          -0.078986
Turbidity                -0.012541
ph_level                 -0.084981
tds_level               -63.213923
chloramines_level        -6.246821
trihalomethanes_level     1.766519
Potability                0.000000
dtype: float64

for col in df_remove_balance.columns:
    plt.figure(figsize=(12, 8))
    sns.histplot(df_remove_balance[col])

df_remove_balance.tds_level.value_counts()

tds_level
1.0    3995
0.0       1
Name: count, dtype: int64

df_remove_balance.drop('tds_level', axis=1, inplace=True)

X, y = df_remove_balance.drop('Potability', axis=1), df_remove_balance['Potability']
X_train, X_test, y_train, y_test = train_test_split(X,y, test_size=0.2, random_state=42)
scale = StandardScaler()
X_train = scale.fit_transform(X_train)
X_test  = scale.transform(X_test)
rfc = RandomForestClassifier(random_state=42)
param_grid_rfc = {
    'n_estimators': [100, 200, 300],
    'max_depth': [None, 10, 20],
    'min_samples_split': [2, 5, 10],
    'min_samples_leaf': [1, 2, 4],
}

grid_search_rfc = GridSearchCV(estimator=rfc, param_grid=param_grid_rfc, verbose=2, n_jobs=-1)
grid_search_rfc.fit(X_train, y_train)
y_pred = grid_search_rfc.predict(X_test)
accur = accuracy_score(y_test, y_pred)
print('Accuracy score: ', accur)
print(classification_report(y_test, y_pred))

Fitting 5 folds for each of 81 candidates, totalling 405 fits
Accuracy score:  0.72125
              precision    recall  f1-score   support

           0       0.72      0.76      0.74       415
           1       0.72      0.68      0.70       385

    accuracy                           0.72       800
   macro avg       0.72      0.72      0.72       800
weighted avg       0.72      0.72      0.72       800

from tensorflow.keras.layers import Dropout
X = df_cleaned.drop('Potability', axis=1)
y = df_cleaned['Potability']
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)
X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.2, random_state=42)
model = Sequential([
    Dense(64, activation='relu', input_shape=(X_train.shape[1], )),
    Dropout(0.5),
    Dense(32, activation='relu'),
    Dropout(0.5),
    Dense(1, activation='sigmoid')
])

model.compile(optimizer='adam',
             loss='binary_crossentropy',
             metrics=['accuracy'])

history = model.fit(X_train, y_train, epochs=50, batch_size=32, validation_split=0.2)
test_loss, test_accuracy = model.evaluate(X_test, y_test)
print("Test Loss:", test_loss)
print("Test Accuracy:", test_accuracy)

C:\Users\Admin\anaconda3\Lib\site-packages\keras\src\layers\core\dense.py:87: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.
  super().__init__(activity_regularizer=activity_regularizer, **kwargs)

Epoch 1/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 14s 16ms/step - accuracy: 0.5062 - loss: 0.7854 - val_accuracy: 0.6030 - val_loss: 0.6770
Epoch 2/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 8ms/step - accuracy: 0.5597 - loss: 0.7183 - val_accuracy: 0.6030 - val_loss: 0.6709
Epoch 3/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.5974 - loss: 0.6865 - val_accuracy: 0.6009 - val_loss: 0.6707
Epoch 4/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.5787 - loss: 0.6804 - val_accuracy: 0.6030 - val_loss: 0.6675
Epoch 5/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.5838 - loss: 0.6858 - val_accuracy: 0.6030 - val_loss: 0.6673
Epoch 6/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 9ms/step - accuracy: 0.6314 - loss: 0.6613 - val_accuracy: 0.6009 - val_loss: 0.6673
Epoch 7/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 13ms/step - accuracy: 0.6198 - loss: 0.6665 - val_accuracy: 0.6009 - val_loss: 0.6663
Epoch 8/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 7ms/step - accuracy: 0.6111 - loss: 0.6733 - val_accuracy: 0.6009 - val_loss: 0.6655
Epoch 9/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 13ms/step - accuracy: 0.6109 - loss: 0.6688 - val_accuracy: 0.6009 - val_loss: 0.6653
Epoch 10/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 13ms/step - accuracy: 0.6229 - loss: 0.6556 - val_accuracy: 0.6030 - val_loss: 0.6644
Epoch 11/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 7ms/step - accuracy: 0.6267 - loss: 0.6553 - val_accuracy: 0.6073 - val_loss: 0.6632
Epoch 12/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 8ms/step - accuracy: 0.6362 - loss: 0.6572 - val_accuracy: 0.6052 - val_loss: 0.6619
Epoch 13/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 7ms/step - accuracy: 0.6129 - loss: 0.6556 - val_accuracy: 0.6094 - val_loss: 0.6594
Epoch 14/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6191 - loss: 0.6611 - val_accuracy: 0.6137 - val_loss: 0.6572
Epoch 15/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6420 - loss: 0.6460 - val_accuracy: 0.6159 - val_loss: 0.6564
Epoch 16/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6338 - loss: 0.6514 - val_accuracy: 0.6223 - val_loss: 0.6546
Epoch 17/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6597 - loss: 0.6376 - val_accuracy: 0.6309 - val_loss: 0.6534
Epoch 18/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6441 - loss: 0.6521 - val_accuracy: 0.6309 - val_loss: 0.6542
Epoch 19/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6327 - loss: 0.6490 - val_accuracy: 0.6330 - val_loss: 0.6520
Epoch 20/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6352 - loss: 0.6397 - val_accuracy: 0.6330 - val_loss: 0.6489
Epoch 21/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6524 - loss: 0.6411 - val_accuracy: 0.6330 - val_loss: 0.6476
Epoch 22/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6425 - loss: 0.6444 - val_accuracy: 0.6245 - val_loss: 0.6464
Epoch 23/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6647 - loss: 0.6266 - val_accuracy: 0.6352 - val_loss: 0.6454
Epoch 24/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6538 - loss: 0.6327 - val_accuracy: 0.6373 - val_loss: 0.6443
Epoch 25/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6640 - loss: 0.6283 - val_accuracy: 0.6438 - val_loss: 0.6418
Epoch 26/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6633 - loss: 0.6311 - val_accuracy: 0.6438 - val_loss: 0.6409
Epoch 27/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6402 - loss: 0.6452 - val_accuracy: 0.6416 - val_loss: 0.6421
Epoch 28/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6558 - loss: 0.6235 - val_accuracy: 0.6373 - val_loss: 0.6396
Epoch 29/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6681 - loss: 0.6338 - val_accuracy: 0.6373 - val_loss: 0.6403
Epoch 30/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 7ms/step - accuracy: 0.6532 - loss: 0.6339 - val_accuracy: 0.6416 - val_loss: 0.6414
Epoch 31/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 8ms/step - accuracy: 0.6407 - loss: 0.6295 - val_accuracy: 0.6395 - val_loss: 0.6404
Epoch 32/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6604 - loss: 0.6210 - val_accuracy: 0.6373 - val_loss: 0.6405
Epoch 33/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6541 - loss: 0.6285 - val_accuracy: 0.6373 - val_loss: 0.6383
Epoch 34/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6704 - loss: 0.6262 - val_accuracy: 0.6330 - val_loss: 0.6367
Epoch 35/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6630 - loss: 0.6238 - val_accuracy: 0.6309 - val_loss: 0.6351
Epoch 36/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6494 - loss: 0.6327 - val_accuracy: 0.6352 - val_loss: 0.6362
Epoch 37/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6581 - loss: 0.6226 - val_accuracy: 0.6309 - val_loss: 0.6364
Epoch 38/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6693 - loss: 0.6143 - val_accuracy: 0.6330 - val_loss: 0.6355
Epoch 39/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 7ms/step - accuracy: 0.6566 - loss: 0.6296 - val_accuracy: 0.6395 - val_loss: 0.6343
Epoch 40/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6593 - loss: 0.6181 - val_accuracy: 0.6309 - val_loss: 0.6338
Epoch 41/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6753 - loss: 0.6152 - val_accuracy: 0.6309 - val_loss: 0.6334
Epoch 42/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6635 - loss: 0.6210 - val_accuracy: 0.6438 - val_loss: 0.6343
Epoch 43/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6696 - loss: 0.6157 - val_accuracy: 0.6395 - val_loss: 0.6359
Epoch 44/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - accuracy: 0.6718 - loss: 0.6161 - val_accuracy: 0.6352 - val_loss: 0.6347
Epoch 45/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6743 - loss: 0.6098 - val_accuracy: 0.6373 - val_loss: 0.6338
Epoch 46/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6772 - loss: 0.6156 - val_accuracy: 0.6373 - val_loss: 0.6346
Epoch 47/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 6ms/step - accuracy: 0.6732 - loss: 0.6192 - val_accuracy: 0.6395 - val_loss: 0.6343
Epoch 48/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 1s 7ms/step - accuracy: 0.6729 - loss: 0.6095 - val_accuracy: 0.6416 - val_loss: 0.6330
Epoch 49/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 7ms/step - accuracy: 0.6596 - loss: 0.6246 - val_accuracy: 0.6438 - val_loss: 0.6328
Epoch 50/50
59/59 ━━━━━━━━━━━━━━━━━━━━ 0s 5ms/step - accuracy: 0.6575 - loss: 0.6375 - val_accuracy: 0.6352 - val_loss: 0.6340
19/19 ━━━━━━━━━━━━━━━━━━━━ 0s 3ms/step - accuracy: 0.6652 - loss: 0.6157 
Test Loss: 0.6122020483016968
Test Accuracy: 0.6752577424049377

df_cleaned

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
0	NaN	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0
1	3.716080	129.422921	18630.057858	6.635246	NaN	592.885359	15.180013	56.329076	4.500656	0
2	8.099124	224.236259	19909.541732	9.275884	NaN	418.606213	16.868637	66.420093	3.055934	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	0
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	0
...	...	...	...	...	...	...	...	...	...	...
3271	4.668102	193.681735	47580.991603	7.166639	359.948574	526.424171	13.894419	66.687695	4.435821	1
3272	7.808856	193.553212	17329.802160	8.061362	NaN	392.449580	19.903225	NaN	2.798243	1
3273	9.419510	175.762646	33155.578218	7.350233	NaN	432.044783	11.039070	69.845400	3.298875	1
3274	5.126763	230.603758	11983.869376	6.303357	NaN	402.883113	11.168946	77.488213	4.708658	1
3275	7.874671	195.102299	17404.177061	7.509306	NaN	327.459760	16.140368	78.698446	2.309149	1

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability
0	7.156857	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0
1	3.716080	129.422921	18630.057858	6.635246	336.094350	592.885359	15.180013	56.329076	4.500656	0
2	8.099124	224.236259	19909.541732	9.275884	330.449166	418.606213	16.868637	66.420093	3.055934	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	0
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	0
...	...	...	...	...	...	...	...	...	...	...
3271	4.668102	193.681735	47580.991603	7.166639	359.948574	526.424171	13.894419	66.687695	4.435821	1
3272	7.808856	193.553212	17329.802160	8.061362	364.091541	392.449580	19.903225	64.327280	2.798243	1
3273	9.419510	175.762646	33155.578218	7.350233	327.357588	432.044783	11.039070	69.845400	3.298875	1
3274	5.126763	230.603758	11983.869376	6.303357	325.952434	402.883113	11.168946	77.488213	4.708658	1
3275	7.874671	195.102299	17404.177061	7.509306	345.728295	327.459760	16.140368	78.698446	2.309149	1

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability	ph_level	tds_level	chloramines_level	trihalomethanes_level
0	7.156857	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0	2	1	1	1
1	3.716080	129.422921	18630.057858	6.635246	336.094350	592.885359	15.180013	56.329076	4.500656	0	0	1	1	0
2	8.099124	224.236259	19909.541732	9.275884	330.449166	418.606213	16.868637	66.420093	3.055934	0	2	1	1	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	0	2	1	1	1
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	0	2	1	1	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3271	4.668102	193.681735	47580.991603	7.166639	359.948574	526.424171	13.894419	66.687695	4.435821	1	0	1	1	0
3272	7.808856	193.553212	17329.802160	8.061362	364.091541	392.449580	19.903225	64.327280	2.798243	1	2	1	1	0
3273	9.419510	175.762646	33155.578218	7.350233	327.357588	432.044783	11.039070	69.845400	3.298875	1	2	1	1	0
3274	5.126763	230.603758	11983.869376	6.303357	325.952434	402.883113	11.168946	77.488213	4.708658	1	0	1	1	0
3275	7.874671	195.102299	17404.177061	7.509306	345.728295	327.459760	16.140368	78.698446	2.309149	1	2	1	1	0

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability	ph_level	tds_level	chloramines_level	trihalomethanes_level
0	7.156857	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0	2	1	1	1
1	3.716080	129.422921	18630.057858	6.635246	336.094350	592.885359	15.180013	56.329076	4.500656	0	0	1	1	0
2	8.099124	224.236259	19909.541732	9.275884	330.449166	418.606213	16.868637	66.420093	3.055934	0	2	1	1	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	0	2	1	1	1
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	0	2	1	1	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3271	4.668102	193.681735	47580.991603	7.166639	359.948574	526.424171	13.894419	66.687695	4.435821	1	0	1	1	0
3272	7.808856	193.553212	17329.802160	8.061362	364.091541	392.449580	19.903225	64.327280	2.798243	1	2	1	1	0
3273	9.419510	175.762646	33155.578218	7.350233	327.357588	432.044783	11.039070	69.845400	3.298875	1	2	1	1	0
3274	5.126763	230.603758	11983.869376	6.303357	325.952434	402.883113	11.168946	77.488213	4.708658	1	0	1	1	0
3275	7.874671	195.102299	17404.177061	7.509306	345.728295	327.459760	16.140368	78.698446	2.309149	1	2	1	1	0

	ph	Hardness	Solids	Chloramines	Sulfate	Conductivity	Organic_carbon	Trihalomethanes	Turbidity	Potability	ph_level	tds_level	chloramines_level	trihalomethanes_level
0	7.156857	204.890455	20791.318981	7.300212	368.516441	564.308654	10.379783	86.990970	2.963135	0	2	1	1	1
1	3.716080	129.422921	18630.057858	6.635246	336.094350	592.885359	15.180013	56.329076	4.500656	0	0	1	1	0
2	8.099124	224.236259	19909.541732	9.275884	330.449166	418.606213	16.868637	66.420093	3.055934	0	2	1	1	0
3	8.316766	214.373394	22018.417441	8.059332	356.886136	363.266516	18.436524	100.341674	4.628771	0	2	1	1	1
4	9.092223	181.101509	17978.986339	6.546600	310.135738	398.410813	11.558279	31.997993	4.075075	0	2	1	1	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3271	4.668102	193.681735	47580.991603	7.166639	359.948574	526.424171	13.894419	66.687695	4.435821	1	0	1	1	0
3272	7.808856	193.553212	17329.802160	8.061362	364.091541	392.449580	19.903225	64.327280	2.798243	1	2	1	1	0
3273	9.419510	175.762646	33155.578218	7.350233	327.357588	432.044783	11.039070	69.845400	3.298875	1	2	1	1	0
3274	5.126763	230.603758	11983.869376	6.303357	325.952434	402.883113	11.168946	77.488213	4.708658	1	0	1	1	0
3275	7.874671	195.102299	17404.177061	7.509306	345.728295	327.459760	16.140368	78.698446	2.309149	1	2	1	1	0

Checking for missing values¶

Dealing with missing data using KNN imputer¶

Perform basic EDA¶

1. Do the histogram to check the distribution and the relationship between all the feature with the target, using hue too¶

3. Checking the correlation and relationship between features and the target using pair plot and heatmap¶

Though there are low correlation between the features and the target, and no clear relationship¶

3. Creating new features¶

Dealing with imbalanced data in the target¶

5 Checking and dealing outliers using boxplot and zscore¶

I have removed outliers, balanced the target and using XGB to check if those methods are optimal. After that i will use KBest and PCA then check for other models and do hyperparameters tuning¶

One mistake that i made is that i didn't scale the features, but the accuracy increase as well as the recall in the class 1 to 0.69 and 0.65 respectively¶

And after removing outliers and balancing the target, the accuracy incrased to 70%. I will try to do KBest and PCA with other models now¶

So i did use KBest, first i tried k=5 which droped the score down to 63 % which is shitty, then i increased the k to 10 and the score increase as well to 69% which is not notable nor good compared to previous score¶

I will try other models: Random Forest¶

So i used random forest without Kbest, removed outliers, balanced the target, and the final for today is 72%¶

I would like to use bagging to check for improvement¶